zalandoresearch/pytorch-vq-vae

Codebook update

GiorgioBarnabo opened this issue · 0 comments

Hi everybody,

looking at the VectorQuantizerEMA nn.Module in the code, I was not able to understand how the codebook vectors are updated after initialization. Is there a way to force the use of all the codebook?
Last, how should I read the perplexity value?

Thank you!

Giorgio